By now everyone’s heard about the potential of artificial intelligence in medicine to revolutionize things like interpreting medical data and predicting patient outcomes. And everyone’s probably heard plenty, too, about how much hype is out there about what these algorithms can actually do.
But what does the evidence say?
Dr. Eric Topol, a cardiologist and geneticist at the Scripps Research Institute in San Diego, recently published a review article in the journal Nature Medicine in which he sifted through the research available on AI in medicine. He wasn’t impressed.
It is awfully hard to get gold-standard results out of AI based on incomplete data. Until AI research databases include patient-generated data on how they experience care and impacts of drugs, devices and other treatments that may be complementary or part of standard-of-care, it is a not sn accurate picture.
Until EHRs include dental data on a par with medical data, AI is analyzing half the picture, and omitting some of the most important and overlooked causative factors. I know this from personal experience. The roots of my escalating health problems were largely dental, and multiple interventions by biological dentists were key to my recovery. None of the granular details are in my EHR, so would not be available to AI.
I have met a growing number of patients like me. We are somewhat incredulous that we recovered from lifelong chronic diseases, but grateful. We are flummoxed that physicians, clinicians, dentists, medical groups, insurers, public agencies, researchers, AI and BigData
seem uninterested and do not pay attention.
When you have a hammer, everything looks like a nail, and you are paid to drive nails, so people count and analyze nails. This is what medicine today feels like for patients with complex chronic illnesses. Connecting the whole body, brain, oral cavity, gut, immune and neurological systems is key to successful diagnosis of underlying causes, planning therapeutic treatments, and achieving positive outcomes.
tog